I limiti dei trial randomizzati in diabetologia

Transcript

1 I limiti dei trial randomizzati in diabetologia Edoardo Mannucci

2 Conflitti di interessi Negli ultimi due anni, E. Mannucci ha ricevuto compensi per relazioni e/o consulenze da: Abbott, AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen, Merck, Novartis, Novo Nordisk, Sanofi, and Takeda. La struttura diretta da E. Mannucci ha ricevuto donazioni, finanziamenti per ricerca o compensi per trial clinici da: AstraZeneca, Bayer, Boehringer Ingelheim, Eli Lilly, Janssen, Merck, Novartis, and Novo Nordisk. 2

3 Cardiovascular risk with insulin therapy in T2DM Observational studies vs RCT Mannucci E, Ferrannini E. Diabetes 3 Obes Metab, 207.

4 Classification of evidence Centre for Evidence-Based Medicine, 4

5 The limitations of randomized trials Selection of population: inclusion and exclusion criteria Selection of population: enrolment of hypercompliant, relatively healthy subjects Jones WS, et al. J Am Coll Cardiol 68: , 5 206

6 From trials to real world: the case of DPP4 inhibitors Phase II-III TECOS AIFA Age (y) Duration DM (y) BMI (kg/m 2 ) HbAc (%) Insulin treatment (%) Prior MACE (%) Monami M, Ahrén B, Dicembrini I, Mannucci E. Diabetes Obes Metab 5:2-20, 203 Green JB et al. N Engl J Med 373: , 205 Montilla S et al. Nutr Metab Cardiovasc Dis6 24:346-53, 204

7 The limitations of randomized trials Selection of population: inclusion and exclusion criteria Selection of population: enrolment of hypercompliant, relatively healthy subjects Specific setting: frequency of visits, recalls, etc. Treatment schedules and dose titration different from routine practice Jones WS, et al. J Am Coll Cardiol 68: , 7 206

8 Degludec: effect on major cardiovascular events Results of the DEVOTE trial Principal endpoint: 3-point MACE (nonfatal MI, nonfatal stroke, and cardiovascular death) 7,637 T2DM patients with prior cardiovascular disease and/or high CV risk, degludec vs glargine U-00 :. Follow-up: 2 y Mean basal insulin dose on trial: 0.6 IU/kg*day Marso SP et al. N Engl J Med 377: , 207 8

9 Exenatide bid vs aspart 30/70 premix: comparison of efficacy Results of a randomized trial Principal endpoint: HbAc at 26 wk Patients: 372 T2DM failing to metformin + SU Treatments: Exenatide bid, BiAsp 30 OADm BiAsp 30 bid all add-on to current treatment Bergenstal R et al. Curr Med Res Opin 25:65-75,

10 Exenatide bid vs aspart 30/70 premix: comparison of efficacy Results of a randomized trial Principal endpoint: HbAc at 52 wk Patients: 50 T2DM failing to combined metformin + SU Treatments: Exenatide bid vs BiAsp 30 bid both add-on to current treatment Nauck MA et al. Diabetologia 50:259-67,

11 Exenatide bid vs aspart 30/70 premix: comparison of efficacy Methods of two randomized trials Bergenstal R et al. Curr Med Res Opin 25:65-75, 2009 Nauck MA et al. Diabetologia 50:259-67, 2007

12 The limitations of randomized trials Selection of population: inclusion and exclusion criteria Selection of population: enrolment of hypercompliant, relatively healthy subjects Specific setting: frequency of visits, recalls, etc. Treatment schedules and dose titration different from routine practice Sample size (inadequate for rare adverse events) Jones WS, et al. J Am Coll Cardiol 68: , 206 2

13 Exenatide LAR: effect on pancreatitis Results of the EXSCEL trial Power calculation: Baseline incidence: 0.09 events/00 pty Relevant increase defined as HR.50 Power 80%, p<0.05 Sample size: 2,752 Holman RR et al. N Engl J Med 377: , 206 3

14 GLPRA and pancreatitis Meta-analysis of randomized trials Power calculation, post-hoc: Relevant increase defined as HR.50 Power 80%, p<0.05 Power: 2% Monami M, Sesti G, Mannucci E, et al. Diabetes Obes Metab 9: 223-4, 207 4

15 The limitations of randomized trials Selection of population: inclusion and exclusion criteria Selection of population: enrolment of hypercompliant, relatively healthy subjects Specific setting: frequency of visits, recalls, etc. Treatment schedules and dose titration different from routine practice Sample size (inadequate for rare adverse events) Duration of trial (inadequate for effects with a long latency) Jones WS, et al. J Am Coll Cardiol 68: , 206 5

16 Glargine insulin: effect on major cardiovascular events Results of the ORIGIN trial ORIGIN trial investigators. N Engl J Med 367:39-28, 202 6

17 Glargine and risk of breast cancer Results of a retrospective cohort study on the UK CPRD Database Risk of breast cancer, glargine vs NPH Women aged>40 y using any basal insulin, in any combination. HR adjusted for confounders (including prior and concurrent treatments, comorbidities, age, etc.) Wu JW et al. J Clin Oncol 35: , 207 7

18 From trials to the real world: pragmatic trials Randomized trials comparing two treatments, with: Inclusion criteria: indications Exclusion criteria: contraindications Procedures: the same as standard clinical practice Jones WS, et al. J Am Coll Cardiol 68: , 206 8

19 Design of a pragmatic trial on glargine U-300 Oster G, et al. Postgrad Med 28:73-739, 206 9

20 The limitations of randomized trials: pragmatic trials Selection of population: inclusion and exclusion criteria Selection of population: enrolment of hypercompliant, relatively healthy subjects Specific setting: frequency of visits, recalls, etc. Treatment schedules and dose titration different from routine practice Sample size (inadequate for rare adverse events) Duration of trial (inadequate for effects with a long latency) Jones WS, et al. J Am Coll Cardiol 68: ,

21 What observational studies on drug treatments can say Verify the characteristics of patients receiving treatment Explore predictors of therapeutic response Extend results of available clinical trials to other populations Explore rare adverse events Explore adverse events with a very long latency (e.g., malignancies) Explore potential drug interactions Compare drugs within the same class (with similar profiles) 2

22 What observational studies on drug treatments cannot say Demonstrate efficacy on HbAc Demonstrate efficacy in preventing cardiovascular disease Substitute randomised trials as a demonstration of cardiovascular safety 2 2